Database schemas are designed in such a way that fields have sufficient space to store the data. However, for most number of instances, the spaces for storing these fields are more than actually used. This, of course, leads to wasted space and resources.
For example, consider a typical database schema for storing customer's address information, as follows:
Table CustmerAddress {Addrline1 varchar(255),AddrLine2 varchar(255),City varchar(100),PinCode varchar(10),Country varchar(50),};
With the above example, a computer database application can ask for a desired number of bytes from a Virtual Memory Manager (e.g., a component of the operating system) for storing a customer's address. However, in this example, only a few (if any) instances of the CustomerAddress Table would use all allocated space, specifically for fields Addrline1 and Addrline2. In such cases, resources are wasted due to the fact that all the allocated spaces for storing the customer's address are not used. These resources (space) can be better implemented to cater to other needs. Also, a considerable amount of energy (i.e., electricity) needs to be consumed to keep up those unused spaces in memory or on external storage devices, as described herein.
A key features of a cloud based paradigm is efficient resource utilization (achieved through virtualization, for example). Many algorithms exist for effective memory allocation usage at a high level. However, there are gaps in the following dimensions:                A data block which is used statically by the process might have the following features: Table “customer” with first name as char 30, middle name as char 10, and last name as char 50. With lack of standards to define data lengths, the designer arbitrarily assigns data lengths and most of the times it is always higher to mitigate the associated risks. In reality the entire reserved space might not be used at all, thus wasting space and resources; and        Commercially Off The Shelf products (COTS) have predefined data tables; ideally the subset domains and respective tables alone will be used. For example, a COTS application for inventory management will have appointment management functionality and related database tables which might not be used at all by the enterprise.        
In all of the above cases, additional energy is used in maintaining the empty unused memory i.e., memory resources are not optimally utilized. In fact, it is known that a modern 1 terabyte drive consumes about 38 microwatts per megabyte. With this metric, 1 Byte would consume 0.036 nanowatt. In a large ecosystem involving multiple applications the percentage of unused memory is more which, in turn, results in the use of considerably more energy consumption.